Synthesis of Site-Specific Radiolabeled Antibodies for

Sep 12, 2016 - Bioconjugate Chem. , 2016, 27 (10), pp 2460–2468 ... To overcome this heterogeneity, we have developed an approach for site-specific ...
6 downloads 0 Views 2MB Size
Article pubs.acs.org/bc

Synthesis of Site-Specific Radiolabeled Antibodies for Radioimmunotherapy via Genetic Code Expansion Yiming Wu,†,∥ Hua Zhu,‡,∥ Bo Zhang,§ Fei Liu,‡ Jingxian Chen,† Yufei Wang,† Yan Wang,† Ziwei Zhang,† Ling Wu,† Longlong Si,† Huan Xu,† Tianzhuo Yao,† Sulong Xiao,† Qing Xia,† Lihe Zhang,† Zhi Yang,*,‡ and Demin Zhou*,† †

State Key Laboratory of Natural and Biomimetic Drugs, School of Pharmaceutical Sciences, Peking University, Beijing 100191, China ‡ Department of Nuclear Medicine, Peking University Cancer Hospital & Institute, Beijing 100142, China § Department of Scientific Research, Peking Union Medical College Hospital, Chinese Academy of Medical Science, Beijing 100073, China S Supporting Information *

ABSTRACT: Radioimmunotherapy (RIT) delivers radioisotopes to antigenexpressing cells via monoantibodies for the imaging of lesions or medical therapy. The chelates are typically conjugated to the antibody through cysteine or lysine residues, resulting in heterogeneous chelate-to-antibody ratios and various conjugation sites. To overcome this heterogeneity, we have developed an approach for site-specific radiolabeling of antibodies by combination of genetic code expansion and click chemistry. As a proof-ofconcept study, model systems including anti-CD20 antibody rituximab, positron-emitting isotope 64Cu, and a newly synthesized bifunctional linker (4-dibenzocyclooctynol−1,4,7,10-tetraazacyclotetradecane-1,4,7,10-tetraacetic acid, DIBO−DOTA) were used. The approach consists of three steps: (1) site-specific incorporation of an azido group-bearing amino acid (NEAK) via the genetic code expansion technique at the defined sites of the antibody as a “chemical handle”; (2) site-specific and quantitative conjugation of bifunctional linkers with the antibodies under a mild condition; and (3) radiolabeling of the chelate-modified antibodies with the appropriate isotope. We used heavy-chain A122NEAK rituximab as proof-of-concept and obtained a homogeneous radioconjugate with precisely two chelates per antibody, incorporated only at the chosen sites. The conjugation did not alter the binding and pharmacokinetics of the rituximab, as indicated by in vitro assays and in vivo PET imaging. We believe our research is a good supplement to the genetic code expansion technique for the development of novel radioimmunoconjugates. antigen binding sites and the Fc domain.6,7 In addition, different drug-to-antibody ratios (DARs) could affect the pharmacokinetics and toxicity of the conjugates.8,9 Therefore, quality control and batch-to-batch reproducibility are often difficult to achieve for heterogeneous products. Site-specific conjugation is one of the most effective techniques for solving the problems mentioned above. In recent years, site-specific modification of the antibodies has been achieved through various chemical and biological strategies, such as introducing extra cysteine residues,10−13 modification of sugars,14−16 and enzymatic conjugation.17−19 Strategies using additional cysteine residues such as the THIOMAB technology are particularly convenient, as the positions of modifications are predetermined and DAR is generally constant. However, these approaches are not without

1. INTRODUCTION Radioimmunotherapy (RIT) is a potent therapeutic technique applicable to numerous cancers. Over the last three decades, due to their remarkable affinity and specificity for cancer biomarkers, antibodies labeled with 124I, 111In, and 64Cu for PET and SPECT imaging and 90Y, 177Lu, and 225Ac for radiotherapy have been successfully developed and translated to the clinic.1,2 A pair of radiolabeled drugs were approved by the FDA:3 the anti-CD20 mAb 90Y-labeled ibrituximab in 20024 and the 131I-labeled tositumomab in 2003.5 Paradoxically, these agents, designed to enable “precision medicine,” are synthesized in a rather imprecise way.6 At present, bifunctional linkers are generally conjugated to monoclonal antibodies (mAbs) through the ε-amino group of lysine or the thiol group of cysteine. However, as antibodies possess approximately 40 lysine and 8 cysteine residues, conventional conjugation produces heterogeneous mixtures with respect to conjugate ratios and sites of conjugation. Biological activities and the plasma half-life of the conjugates may decrease, respectively, with modifications at the © 2016 American Chemical Society

Received: July 22, 2016 Revised: September 1, 2016 Published: September 12, 2016 2460

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry

wild type rituximab reached ∼20 mg/L in the shake flask. Correct assembling of the antibody was confirmed by nonreducing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), with the whole antibody appearing at 160 kDa, while heavy and light chains appeared at 55 and 24 kDa, respectively (Figure 1A). To site-specifically incorporate NEAK into the anti-CD20 antibody, an orthogonal amber suppressor tRNA−aaRS pair,34

limitations,13 including the potential formation of unwanted disulfide bonds that may lead to protein misfolding. In addition, absolute site-specific conjugation is difficult to achieve because relatively nonspecific reduction−oxidation steps are required. Furthermore, cysteine-mediated maleimide-conjugated linker is intrinsically unstable,20,21 which increases the risk of off-target nuclides. In response to these issues, extensive efforts have been made to develop robust and reliable site-selective modification methodologies for antibodies. Among these, the combination of the genetic code expansion and bio-orthogonal “click” chemistry proved to be a successful strategy for conjugating molecules in a highly specific and reproducible manner.7,21−25 The resulting conjugates possessed excellent pharmacokinetics and potent in vitro cytotoxic activity and achieved complete tumor regression in a xenograft study.7 In this study, using an anti-CD20 antibody rituximab as an example, we explored the potential of the approach described above for site-specific antibody-nuclides labeling. The procedure we used consists of three key steps: (1) site-specific incorporation of an azido group-bearing amino acid, Nε-2azideoethyloxycarbonyl-L-lysine (NEAK), via the genetic code expansion technique at the defined sites on the antibody as a chemical handle;26 (2) site-specific and quantitative conjugation of bifunctional linkers with antibodies under mild conditions; and (3) radiolabeling of the chelate-modified antibodies with the appropriate isotope. Using heavy-chain A122NEAK rituximab, we obtained homogeneous radioconjugates with precisely two chelates per antibody, incorporated only at the chosen sites. We used in vitro assays and in vivo PET imaging to evaluate the effects of the conjugation on antigen binding and pharmacokinetics of rituximab.

2. RESULTS 2.1. System Design. To achieve site-specific labeling with radioisotopes, rituximab was chosen as the model antibody and labeled with 64Cu and 177Lu. As for the bifunctional linker, 1,4,7,10-tetraazacyclotetradecane-1,4,7,10-tetraacetic acid (DOTA) was used to chelate the nuclides because of its stability.27,28 DOTA has already shown good biological performance when used in protein conjugation of various radioisotopes.29,30 Rituximab is a chimeric monoclonal antibody (mAb) targeting the CD20 antigen and has been approved by the United States Food and Drug Administration for the treatment of non-Hodgkin lymphomas (NHL).31 The CD20 antigen is a hydrophobic transmembrane protein, which is expressed on B-lymphocytes. Furthermore, the CD20 antigen is exposed at the cell surface without any free isoform in the serum; no internalization happens following rituximab binding. These properties make the CD20 receptor a good candidate for radioimmunotherapy. Several studies of radiolabeled anti-CD20 monoclonal antibodies with nonmyeloablative doses have been reported in the treatment of B-cell NHL.30,32,33The most commonly used nuclides for radioimmuno-therapy are 131I, 90Y, 68Ga, 64Cu, and 177Lu.27 In this study, 64Cu and 177Lu were chosen for radiolabeling and positron emission tomography (PET) imaging because they are well-characterized and easily obtained. 2.2. Antibody Expression and Site-Specific Incorporation of the Azide-Bearing Unnatural Amino Acid. The suspension 293 mammalian cell expression system (FreeStyle 293, Life Technology Inc.) was chosen to express the intact antibody with the mature post-translational modification. The conditions of fermentation were optimized, and the yield of

Figure 1. Site-specific incorporation of NEAK into rituximab. (A) SDS-PAGE gel of the WT rituximab and HC-A122NEAK rituximab with or without DTT. Both heavy chains of the reduced antibody migrated at 55 kDa, and the light chains migrated at 24 kDa. (B) The ncAA incorporation sites on the Fab of rituximab. The antibody (PDB: 2OSL) was drawn in salmon cartoon mode, and the mutated residues, H-A122 and L-K168, were shown as cyan sticks with the surface indicated. (C) Western blot analysis of NEAK-dependent expression of rituximab (with 6× His tag at the C terminus of the heavy chain) using the anti-His antibody. A series of sites chosen to NEAK incorporation was shown. (D) Verification of NEAK incorporation at the defined site (H-359R chosen as a representative site) by LC−MS/ MS peptide sequencing. (E) The HPLC analysis of the purified HCA122NEAK rituximab showed the retention time at 16.2 min, similar to the BSA dimer (∼134 kDa). (F) Flow cytometry analysis of the HC-A122NEAK rituximab antigen binding. Raji cells and Ramos RA1 cells were used as CD20-positive cells, and Jurkat cells were used as CD20-negative cells. 2461

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry

A122NEAK, ∼1 mg/mL) was conjugated with the newly synthesized DIBO−DOTA linker (2 mM) in PBS buffer (conjugation buffer) overnight at 4 °C (12 h). Buffer was exchanged with the radiolabeling buffer using the 50 kDa centrifugal filter unit, and additional washing was done to remove the excess linker. Conjugation of the RTX-WT was performed in the same way except for using HEPES (pH 8.5) as the conjugation buffer. All buffers including samples were pretreated with Chelax-100 resin to remove residual metal ions. To assess the ratio of chelates per antibody, ESI intact mass analysis was performed. Conventional conjugation of the S-2(4-isothiocyanatobenzyl)-1,4,7,10-tetraazacyclododecane tetraacetic acid (DOTA−SCN) to lysines resulted in a heterogeneous mixture with number of linkers per antibody ranging from 5 to 7 (Figure 2A), which was consistent with previously reported.37 In contrast, site-specific conjugation of the DIBO− DOTA to the azide-bearing rituximab showed an exact molecular-weight shift of two bifunctional linkers compared with that of the unconjugated antibody (Figure 2B), which confirmed the homogeneous distribution of two linkers per antibody. The analyses of reduced and deglycosylated samples were also conducted. The conventional conjugation of p-SCN-BnDOTA to lysines resulted in a heterogeneous mixture with an average of 0−1 DOTA on the light chain and 1−4 DOTA on the heavy chain (Figure S3A). In contrast, site-specific conjugation, in which NEAK was incorporated at two specific locations on the heavy chain, showed a homogeneous distribution of 1 DOTA per heavy chain with no molecularweight shift appearing on the light chain (Figure S3B). Thus, homogeneous conjugates with two bifunctional linkers at the specific sites were produced through genetically encoded unnatural amino acids. The surface plasmon resonance (SPR) assay was performed to evaluate the effects of conjugation on the antibody antigen binding. No significant difference was found between the sitespecific conjugated and unconjugated rituximab (Figure 2C). In contrast, the equilibrium constant of rituximab−CD20 binding decreased in nonselective conjugation group, owing to the decrease of binding ability but not dissociation. 2.4. Radiolabeling with 64Cu or 177Lu. The purified rituximab conjugates (site-specific and nonselective) were labeled with 64Cu or 177Lu at 42 °C. The process of labeling was monitored using instant thin-layer chromatography (ITLC) after chelating the reaction mixture with EDTA, and the radiolabeling efficiency was determined using both ITLC and size-exclusion high-performance liquid chromatography (SE-HPLC). Our results revealed that >95% labeling efficiency was achieved for 64Cu and 177Lu labeling; the radiochemical purity of the radioconjugates was up to 99% after PD-10 purification (Figure 3A,B); their specific activity was 280 MBq/ nmol. SDS-PAGE of the labeled conjugates, followed by Coomassieblue staining and radiography, was performed to further validate the site-specific conjugation. Our results showed that labeling with conventional conjugation was nonselective, occurring on both the heavy and the light chains of the antibody (Figures 3C and S2C). In contrast, in labeling with RTX-HC-A122NEAK, only the heavy chain was labeled with the radioisotope, further verifying the site-specificity and stoichiometry of the method (Figures 3C and S2C). The in vitro stabilities of the 64Cu-DOTA−rituximab conjugates, in both PBS and serum, were tested as depicted in Figure S2D. Excellent stabilities were observed because more than 95%

derived from the corresponding Methanosarcina mazei, was coexpressed with mutated rituximab genes containing a TAG codon at residue K168 (LC-K168NEAK) on the light chain or A122 (HC-A122NEAK), K278, E322, R359, and K396 on the heavy chain (Figure 1B). These sites were chosen on the basis of their solvent accessibility and distance from the active site.7,35 A strong band was detected by Western blotting with the NEAK in the culture medium, and no expression was found without NEAK (Figure 1C), which supported the suitability of genetic code expansion in engineering NEAK to rituximab. Incorporation of NEAK at the specific site was confirmed by peptide sequencing (Figure 1D). The mutant anti-CD20-IgG was characterized by nonreducing SDS-PAGE and showed a band at 160 kDa. Reducing conditions revealed a light chain band at 24 kDa and a heavy chain band at 55 kDa, similar to the wild type rituximab (Figure 1A). The yields of all the mutants were ∼2 mg/L when grown in shake flasks after nickel nitrilotriacetic acid (Ni-NTA) affinity purification. To avoid disturbing antigen−antibody binding and functions driven by the Fc domain, K168LC and A122HC were chosen for further experiments (Figure 1B). HPLC analysis demonstrated that high purity was achieved (>99%) with the retention time of 16.2 min for the mutant antibody (Figure 1E). The binding ability of the mutants to the CD20 extracellular domain was indistinguishable from wild type (WT) rituximab, as determined by flow cytometry and immunofluorescence staining (Figures 1F and S1D) using Raji and Ramos RA1 cells as the CD20+ cells and Jurkat cells as the negative control. 2.3. Preparation and Characterization of the Rituximab−DOTA Conjugates. To label the antibody with radioisotope site-specifically, a bifunctional linker (DIBO− DOTA) was synthesized with DOTA on one end for chelating the nuclides and a cyclooctyne group (4-dibenzocyclooctynol, DIBO) on the other end for site-specific conjugation of the azide-bearing antibody via copper free [3 + 2] cycloaddition click chemistry.36 The synthesis is presented in Scheme 1. Rituximab containing NEAK at position HC-A122 (RTX-HCScheme 1. (A) Synthesis of the DIBO−DOTA bifunctional linker. (B) The structure of NEAK

2462

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry

Figure 2. Determination of the chelate-to-antibody ratio via intact mass. (A) The electrospray ionization mass spectrometry (ESI-MS) of the native WT rituximab with the molecular weight of 147263.66 Da (left); the ESI-MS of the WT native rituximab conjugated with 5, 6, and 7 p-SCN-BnDOTA, showing the molecular weight of 150002.46, 150555.33, and 151106.78 Da, respectively (right). (B) The ESI-MS of the native HCA122NEAK rituximab with the molecular weight of 149488.89 Da (left); the ESI-MS of the native HC-A122NEAK rituximab conjugated with DIBO−DOTA site-specifically with the molecular weight of 151206.38 Da (right). (C) Surface plasmon resonance sensorgrams of the WT rituximab, lysine-conjugated rituximab, and site-specific conjugated rituximab binding to immobilized MBP-CD20. Antibodies were injected over the sensor surface at five different concentrations (100, 33.3, 11.1, 3.7, and 1.2 nM).

conjugates, upon incubation at 37 °C up to 60 h, retained their original structures. 2.5. Small-Animal PET Imaging and in Vivo Pharmacokinetics. The huCD20TM model, mimicking the human Bcell lymphoma, was used to evaluate the in vivo behavior of 64 Cu site-specific labeled antibody. In vivo targeting ability of the 64Cu-DIBO−DOTA−RTX-A122NEAK in huCD20TM was demonstrated at various time points after tail-vein injection of the radiopharmaceutical, with and without blocking of the CD20 antigens by cold rituximab. Each mouse was imaged at various time points (0, 18, and 60 h) after a tracer injection of ∼1 mCi of radioconjugates. For both groups, the majority of the activity was in the tumor, blood, and liver at both time points. Our results reveal that the uptake in the CD20-positive tumors was significantly higher than in the blocked control. No difference, except in the liver and spleen, was observed in background tissue uptake between the two groups (Figure 4A). Images show that the radioactivity in blood, kidney, and heart decreased between 18 and 60 h measurements, and it remained

constant in the liver and the spleen. Thus, we concluded that the uptake of the site-specific radioconjugate occurred primarily in organs with high CD20 expression. To further explore the pharmacokinetics of the conjugates in vivo, ELISA was performed to analyze the concentration changes of serum rituximab. WT rituximab, HC A122NEAK rituximab, and two conjugates (the WT-Rituximab with p-SCNBn-DOTA conjugated to lysines and HC A122NEAK rituximab with DIBO−DOTA conjugated to NEAK) were tested, with three male CB-17 SCID mice in each group. Each mouse was intravenously injected with the sample (10 μg), and the serum was collected at 0, 6, 24, 48, and 72 h after injection. The rituximab concentration was quantified using an ELISA kit against the Fc domain of human IgG. Similar pharmacokinetics of the four samples indicated that NEAK and conjugation had no significant effect on pharmacokinetics of the antibody (Figure 4B). 2463

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry

3. DISCUSSION For the past two decades, substantial efforts have been made in the development of therapeutic mAbs for the treatment of a wide range of diseases, including cancers and infectious and immunological diseases. Because of their high binding affinity and specificity, mAbs are ideal carriers for delivery of various cytotoxic molecules to therapeutic target sites, including cytotoxic drugs or radionuclides. Currently available antibodydrug conjugations (ADCs) are heterogeneous, with zero to eight drug molecules per antibody, which can influence the pharmacokinetics and in vivo performance of the formulation.6 Recently, the genetic code expansion has become a promising tool for protein drug design.7,8,23,26 This approach was successfully applied to site-specific PEGylation,26 virus labeling,38 and ADCs design.22 On the basis of the alkoxyamine group of an unnatural amino acid and orthogonal chemical reactions, Axup et al. (2012)7 demonstrated improved efficacy, specificity, and stability of site-specific auristatin conjugates of trastuzumab (a monoclonal anti-Her2 antibody). In this study, azide-bearing unnatural amino acid NEAK was incorporated genetically at defined sites of the antibody as the “chemical handle” (Figure 1C). Bifunctional linker DIBO− DOTA was synthesized and conjugated with the encoded NEAK amino acid through the cyclooctyne-azide click reaction. In line with the orthogonality and stoichiometry of the reaction, no detectable conjugation of native residues occurred, resulting in exactly two bifunctional molecules per antibody (Figures S2B, S2C, and 3C). Notably, our results indicated that no unreacted antibody was observed after conjugation (Figures 2B and S3B), suggesting >95% coupling efficiency achieved under mild conjugation conditions (4 °C overnight). The comparison of the in vitro and in vivo behavior of coupled and uncoupled antibodies indicated that comparable properties were obtained between site-specific conjugates and unconjugated IgG. In contrast, nonselective conjugation significantly affected the binding capacity of the antibody (Figure 2C). The nuclides 64 Cu or 177Lu were used to label the antibody through the conjugated bifunctional linker. The fidelity of labeling was confirmed by SDS-PAGE and radiography (Figures S2B, S2C, and 3C), and the biodistribution of radioconjugates was measured by in vivo PET imaging. According to our results (Figure 4A), high blood-pool activity and background uptake were evident at early time points, but over the course of the experiment, the tumor signal increased significantly to a point at which it is by far the most prominent feature in the image. Together with the pharmacokinetic analysis described above, these data suggested favorable properties of the site-specific radioconjugate in vivo. The generation of antibody small-molecule conjugate, although sounding straightforward, is very complicated in practice. Continuous studies have been ongoing for exploration of site-specific antibodies−drug conjugation. The very recent work by Zeglis et al. reported an enzymatic transformation approach by which PEGylated TCO moieties were appended, using a click chemistry strategy, to the heavy-chain glycans of huA33 antibody.39 The enzymatic approach led N-glycans to be modified, which, due to the potential heterogeneity of the Nglycans, might generate heterogeneous conjugates. As a contrast, we reported the site-specific labeling of antibodies with radioisotopes through the use of an unnatural amino acid (Uaa). The azide-bearing Uaa NEAK was incorporated into the heavy chain of rituximab by reassignment

Figure 3. Radiolabeling of DOTA−rituximab and the analysis of sitespecific labeling. (A) Radiochemical purity of the 64Cu site-specific labeled conjugates before (left) and after (right) PD-10 purification, as tested by ITLC. (B) Radiochemical purity of the 64Cu site-specific labeled conjugates after PD-10 purification, as tested by SE-HPLC. (C) SDS-PAGE radiography of the WT rituximab and HCA122NEAK rituximab. Only the heavy chain is labeled with 64Cu on the HC-A122NEAK rituximab, and both of the chains are labeled with 64 Cu on the WT rituximab.

Figure 4. Small-animal PET imaging and in vivo pharmacokinetics. (A) PET imaging of the Ramos RA1 tumor-bearing mice at 18 and 60 h post-injection of the site-specifically conjugated rituximab. The blocked controls were injected with 500 μg of WT rituximab. (B) Pharmacokinetics analysis of unconjugated and conjugated rituximab in CB-17 SCID mice. Each mouse was administered 10 μg of the antibody by intraveous injection, and the blood concentration of the rituximab was monitored using an anti-human IgG1 ELISA kit. 2464

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry

4.3. Antibody Expression and Purification. The genes encoding the heavy and light chain of rituximab were synthesized by Life Technology Inc. and cloned into the vector pEF1/Myc-His b (Life) within sites of KpnI and NotI (Figure S1B). A 6× His tag was added at the C-terminal for detection and affinity purification. For antibody expression (RTX-WT), the suspension and transient mammalian expression system, Freestyle 293, was used according to manufacturer’s instructions (Figure S1A). After 7 days of expression, the supernatant was collected and concentrated (Figure S1C). The protein was purified by Ni-NTA His-Bind Resin (Novagen) and Superdex200 Increase Column (GE). Main elution peak was collected, concentrated, and bufferexchanged into PBS buffer using a 50 kDa centrifugal filter unit (Millipore). To generate the antibody with azido modification, the UAG (amber) stop codon was introduced to the conjugation sites, where unnatural amino acid would be incorporated with the aid of engineered orthogonal tRNA and aminoacyl tRNA synthetase pairs. The plasmids encoding NEAK-containing RTX variants were generated by site-directed PCR mutagenesis. The obtained mutant plasmid, pEF1b-RTX-TAG, was transfected into 293F cells with pCMV-mmPylRS−tRNA CUA plasmid. Expression of NEAK-bearing rituximab variants was similar to the expression of RTX-WT except for the addition of 1 mM NEAK in the medium. 4.4. Preparation of DIBO−DOTA and DOTA−SCN Antibody Conjugates. The bifunctional linker DIBO− DOTA was conjugated to the antibody as follows: 1 to 5 mg/mL NEAK-bearing antibody in PBS was conjugated with a 20-fold molar excess of DIBO−DOTA overnight at 4 °C. The conjugates were desalted to remove the unconjugated DIBO− DOTA using the 50 kDa centrifugal filter unit (Millipore) and exchanged to the isotope-labeling buffer (0.2 mol/L ammonium acetate buffer, pH 7.0). Metal ions were removed from the conjugation and labeling buffer before use via the Chelex 100 resin (BioRad). Protein concentration was determined by spectrophotometry at 280 nm. The conjugation of DOTA−SCN with the RTX-WT was conducted as previously described.40,41 Briefly, the antibody was buffer-exchanged to the 0.2 M HEPES buffer (pH 8.5) with the final concentration about 10 mg/mL. The RTX-WT was then incubated with DOTA-SCN at 37 °C for 2 h in molar ratios of 1:10. The conjugates were buffer-exchanged into the metal-free radiolabeling buffer. 4.5. In Vitro Stability Test of 64Cu−DOTA−Rituximab. The stability of 64Cu-DOTA−rituximab was examined by measuring its radiochemical purity (RCP) by performing radio high-HPLC at each time points after the purification. In short, 64 Cu-DOTA−rituximab (3.7 MBq) was added to the test tubes containing 0.1 M PBS (pH 7.4) and 5% BSA, respectively. The mixture was incubated under shaking conditions at 37 °C in a thermomixer. The RCP was measured at 1, 8, 24, 48, and 60 h by radio-HPLC. 4.6. Determination of the Chelate-to-Antibody Ratio. The chelate-to-antibody ratio was analyzed with ESI-MS. Briefly, the intact mass of the conjugated and unconjugated antibodies was determined on a Dionex 3000 UHPLC separation system in tandem with an Orbitrap Velos (Thermo Scientific, West Palm Beach, FL). Samples were applied onto a POROS R2 10 μm column, 2.1 × 100 mm (Applied Biosystems, Carlsbad, CA). The column was equilibrated in 75% mobile phase A (0.1% TFA in water) and 25% mobile

of stop codon TAG as a sense codon, conjugated with a bifunctional linker through the Uaa, and then radiolabeled via the chelator on the linker. Theoretically, any position at the antibody could be chosen for conjugation by this way. Further thorough and systematic research is required to examine the advantages of site-specific radiolabeling. Previous studies have shown significant advantages of site-specific conjugation on activities, stability, and pharmacokinetics.9,15,21,24,37 The biggest benefit of our site-specific conjugation method is the controllability of drug load location and drug-to-antibody ratio, which could improve batch-to-batch consistency of the conjugates and avoid the potential risks of nonselective conjugation, including a decrease in antigen affinity if coupling occurs at the antigen-binding domains, or changes in pharmacokinetics when coupling to FcRn domains. In summary, we have developed a method for site-selective and -specific radiolabeling of antibodies by a combination of genetic code expansion and catalyst-free click chemistry. To our knowledge, this is one of very few reports on site-specific radioisotope labeling of antibodies. As for this proof-of-concept system, the site-specific radioimmunoconjugates showed improved homogeneity without changing the pharmacokinetics and antigen binding of the antibody. We believe that this strategy is a good supplement to the genetic code expansion application to antibody−drug conjugates, which could play a critical role in the development of novel well-defined and highly selective radioimmunoconjugates for laboratory and clinic.

4. EXPERIMENTAL PROCEDURES 4.1. General Materials. The plasmid pCMV-mmPylRS− tRNACUA used for incorporation of UAA was kindly donated by Prof. Peng Chen. The Jurkat, Raji, and Ramos RA1 cell lines were purchased from the ATCC, and Freestyle 293 and Freestyle CHO-S suspension-adapted cells were bought from Invitrogen. PEI (25 kDa) and PEI (40 kDa) were purchased from Polysciences. The 4-dibenzocyclooctynol (DIBO) and Nε-2-[(azidoethoxy)carbonyl]-L-lysine (NEAK) were synthesized as previously reported. The commercial rituximab (RTX-WT) used for conjugating with p-SCN-Bn-DOTA (DOTA−SCN, Macrocyclics Inc.) was purchased from Roche. 4.2. Synthesis of the Bifunctional Linker DIBO−DOTA. After ethylenediamine (30 mg, 0.5 mmol) and compound 2 (Figure S4) (38 mg, 0.1 mmol) were dissolved in DCM (10 mL), 300 μL of triethylamine was added and stirred overnight at room temperature. The reaction mixture was evaporated to dryness and purified by silica gel column chromatography using DCM−MeOH (10:1), yielding the N-(2-aminoethyl)-11,12didehydro-5,6-dihydrodibenzo[a,e]cycloocten-5-yl ester as yellow solid. Compound 3 (31 mg, 0.1 mmol) was dissolved in 5 mL of DMF, and DIEA (65 mg, 0.5 mmol) and p-SCN-BnDOTA (55 mg, 0.1 mmol) were added. The resulting reaction mixture was stirred at room temperature for 1 h, by which time reaction was deemed complete by HPLC analysis. The compound was purified by preparative HPLC, yielding the desired product as an off-white solid (Figure S4). 1H NMR (300 MHz, D2O): 9.63 (s, 1H, CONH), 8.14−7.09 (12H, aromatics), 5.30 (s, 1H, CHOH), 3.57−2.33 (31H). 13C NMR (75 MHz, CDCl3): 181.5, 164.0, 163.2, 156.4, 153.3, 151.8, 131.2, 130.4, 130.0, 129.4, 129.3, 128.9, 128.6, 128.3, 128.2, 127.0, 126.7, 124.9, 123.8, 121.3, 113.5, 110.8, 76.4, 65.8, 46.5, 44.4, 36.7, 31.7. MALDI HRMS (m/z): [M + H+] calcd for C23H16NO5+: 858.3452; found, 858.3496. 2465

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry

the PET scan. Static scans (1−20 min) were obtained, and the images were reconstructed using a two-dimensional ordered subsets expectation maximization (OSEM) algorithm. All animal experiments were performed in accordance with guidelines of the Animal Care and Use Committee of Peking University (Ethics Approval License: 2015KT08).

phase B (0.1% TFA in 90% acetonitrile and 10% HPLC H2O) with a gradient of 2.5%/min over 10 min with a flow rate of 0.3 mL/min. Column temperature was set to 70 °C. Intact mass spectra were collected in positive ion mode with the mass range set to 700−3000 m/z with capillary and source temperatures of 300 °C, a sheet gas flow rate of 50 arbitrary units, and an auxiliary gas flow rate of 8 arbitrary units. The capillary voltage was set to 3500 V with the fragmentor voltage set to 75 V. The acquired spectra were deconvoluted using Thermo Deconvolution 2.0 software. Reduced and deglycosylated samples were used for further analysis. Briefly, the antibodies were treated with PNGase overnight at 37 °C in 0.1 M Tris−HCl (pH 8.0), followed by reduction in 4 M guanidine−HCl, 0.1 M Tris−HCl (pH 8.0), and 0.05 M DTT for 10 min at 85 °C. Other analysis procedures were similar to procedures described above. 4.7. Radiolabeling of DOTA−Rituximab and the Analysis of Site-Specific Labeling. The radiolabeling of DOTA−rituximab was carried out with 64CuCl2 (provided by Peking University Cancer Hospital and Institute, Beijing, China). DOTA−DIBO−rituximab and DOTA−SCN−rituximab (100−150 μg/mL in 0.2 M ammonium acetate buffer (pH 5.5)) were reacted with 250 μCi 64CuCl2 solution at 42 °C for 1 h. The labeling procedure was monitored with ITLC and radio-HPLC as previously reported. After incubation, 64CuDOTA−rituximab was purified by size-exclusion chromatography using a PD-10 column. The radiolabeling of DOTA− rituximab with 177Lu was similar except that the pH of reaction buffer was 7.0 and the incubation time was longer (∼2 h). The results of site-specific radiolabeling were inspected by radioautography. 4.8. Flow Cytometry. Jurkat, Raji or Ramos RA1 cells (1 ×107 ) were washed with cold PBS twice and incubated with 4% PFA for 0.5 h. The suspension was discarded and the cells cultured with NEAK-containing rituximab (10 μg/mL) and 0.1% Triton PBS for 1 h. The cells were washed with cold 0.1% Triton PBS three times and cultured with anti-human Fc monoantibody (5 μg/mL) labeled with Alexa-488 for 1 h in the dark. The cells were washed with cold PBS three times and resuspended using 100 μL of PBS. Analysis was carried out on a FACSscan cytometer using CELLQUEST software. 4.9. Surface Plasmon Resonance Analysis. Receptor binding affinities were measured by BIAcore 3000 (GE) using the CM5 sensor. The major extracellular loop of CD20 was cloned into the pMalc5x vector, and MBP/MBP-CD20 were expressed in BL21(DE3), as previously reported (Figure S2A). Analysis of the interactions between MBP−MBP-CD20 and rituximab was performed in HBS buffer (10 mM HEPES, 3 mM EDTA, and 150 mM NaCl; pH 7.4) at a flow rate of 30 μL/min. The MBP−MBP-CD20 was directly immobilized onto the surface of the sensor at pH 5.0. All samples were measured at five different concentrations ranging from 1.2 to 100 nM by serial 3-fold dilutions. Between the measurements, regeneration was performed in 10 mM NaOH buffer (30 μL). The experimental data were fitted and analyzed using the BIA evaluation software by a simple one-to-one kinetic model. 4.10. Small-Animal PET Imaging. The PET imaging was carried out on a microPET rodent model scanner as described earlier.30 Prior to imaging experiments, Ramos RA1 tumorbearing SCID CB17 mice were administered 64Cu-DOTA− rituximab (0.5 mCi) via lateral tail veins. At each time point (0, 18, and 60 h), the animals were anesthetized with 2% isoflurane at room temperature and placed in the appropriate position for



ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.bioconjchem.6b00412. Figures showing optimization of the rituximab transient transfection conditions; SDS-PAGE of the purified MBP and MBP-CD20 for the SPR assay; SDS-PAGE of the WT rituximab, HC-A122NEAK rituximab, and LCK168NEAK rituximab; SDS-PAGE radiography of the WT and HC-A122NEAK rituximab; determination of the chelate-to-antibody ratio via intact mass; and 1HNMR and 13C-NMR spectra of compounds 2 and 4. (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] *E-mail: [email protected]. Author Contributions ∥

These authors contributed equally to this work.

Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This research was financially supported by the National Natural Science Foundation of China (grant no. 81530090) and National Basic Research Program of China (grant nos. 2016YFA0501500 and 2010CB12300).



ABBREVIATIONS RIT, radioimmunotherapy; NEAK, Nε-2-azideoethyloxycarbonyl−L-lysine; mAbs, monoclonal antibodies; DARs, drugto-antibody ratios; DIBO, 4-dibenzocyclooctynol; PET, positron emission tomography; DOTA, 1,4,7,10-tetraazacyclotetradecane-1,4,7,10-tetraacetic acid; NHL, non-Hodgkin lymphomas; ADCs, antibody−drug conjugations; RCP, radiochemical purity; ESI-MS, electrospray ionization mass spectrometry; SPR, surface plasmon resonance; OSEM, ordered subsets expectation maximization



REFERENCES

(1) Wu, A. M. (2009) Antibodies and antimatter: the resurgence of immuno-PET. J. Nucl. Med. 50 (1), 2−5. (2) van Dongen, G. A., Visser, G. W., Lub-de Hooge, M. N., de Vries, E. G., and Perk, L. R. (2007) Immuno-PET: a navigator in monoclonal antibody development and applications. Oncologist 12 (12), 1379− 1389. (3) Macklis, R. M. (2007) Radioimmunotherapy as a therapeutic option for Non-Hodgkin’s lymphoma. Semin Radiat Oncol 17 (3), 176−183. (4) Mondello, P., Cuzzocrea, S., Navarra, M., and Mian, M. (2016) 90 Y-ibritumomab tiuxetan: a nearly forgotten opportunity. Oncotarget 7 (7), 7597−7609. (5) William, B. M., and Bierman, P. J. (2010) I-131 tositumomab. Expert Opin. Biol. Ther. 10 (8), 1271−1278.

2466

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry

Part 2: Peptide Tags and Unnatural Amino Acids. Mol. Imaging Biol. 18 (2), 153−165. (23) Tian, F., Lu, Y., Manibusan, A., Sellers, A., Tran, H., Sun, Y., Phuong, T., Barnett, R., Hehli, B., Song, F., et al. (2014) A general approach to site-specific antibody drug conjugates. Proc. Natl. Acad. Sci. U. S. A. 111 (5), 1766−1771. (24) Kularatne, S. A., Deshmukh, V., Ma, J., Tardif, V., Lim, R. K., Pugh, H. M., Sun, Y., Manibusan, A., Sellers, A. J., Barnett, R. S., et al. (2014) A CXCR4-targeted site-specific antibody-drug conjugate. Angew. Chem., Int. Ed. 53 (44), 11863−11867. (25) Lim, R. K., Yu, S., Cheng, B., Li, S., Kim, N. J., Cao, Y., Chi, V., Kim, J. Y., Chatterjee, A. K., Schultz, P. G., et al. (2015) Targeted Delivery of LXR Agonist Using a Site-Specific Antibody-Drug Conjugate. Bioconjugate Chem. 26 (11), 2216−2222. (26) Zhang, B., Xu, H., Chen, J., Zheng, Y., Wu, Y., Si, L., Wu, L., Zhang, C., Xia, G., Zhang, L., et al. (2015) Development of next generation of therapeutic IFN-alpha2b via genetic code expansion. Acta Biomater. 19, 100−111. (27) Sarko, D., Eisenhut, M., Haberkorn, U., and Mier, W. (2012) Bifunctional chelators in the design and application of radiopharmaceuticals for oncological diseases. Curr. Med. Chem. 19 (17), 2667− 2688. (28) Panwar, P., Iznaga-Escobar, N., Mishra, P., Shrivastava, V., Sharma, R. K., Chandra, R., and Mishra, A. K. (2005) Radiolabeling and biological evaluation of DOTA-Ph-Al derivative conjugated to anti-EGFR antibody ior egf/r3 for targeted tumor imaging and therapy. Cancer Biol. Ther. 4 (8), 854−860. (29) Zacchetti, A., Coliva, A., Luison, E., Seregni, E., Bombardieri, E., Giussani, A., Figini, M., and Canevari, S. (2009) 177)Lu- labeled MOv18 as compared to (131)I- or (90)Y-labeled MOv18 has the better therapeutic effect in eradication of alpha folate receptorexpressing tumor xenografts. Nucl. Med. Biol. 36 (7), 759−770. (30) Natarajan, A., Gowrishankar, G., Nielsen, C. H., Wang, S., Iagaru, A., Goris, M. L., and Gambhir, S. S. (2012) Positron emission tomography of 64Cu-DOTA-Rituximab in a transgenic mouse model expressing human CD20 for clinical translation to image NHL. Mol. Imaging Biol. 14 (5), 608−616. (31) McLaughlin, P. (2001) Rituximab: perspective on single agent experience, and future directions in combination trials. Crit Rev. Oncol Hematol 40 (1), 3−16. (32) Natarajan, A., Habte, F., and Gambhir, S. S. (2012) Development of a novel long-lived immunoPET tracer for monitoring lymphoma therapy in a humanized transgenic mouse model. Bioconjugate Chem. 23 (6), 1221−1229. (33) Forrer, F., Chen, J., Fani, M., Powell, P., Lohri, A., Muller-Brand, J., Moldenhauer, G., and Maecke, H. R. (2009) In vitro characterization of (177)Lu-radiolabelled chimeric anti-CD20 monoclonal antibody and a preliminary dosimetry study. Eur. J. Nucl. Med. Mol. Imaging 36 (9), 1443−1452. (34) Nguyen, D. P., Lusic, H., Neumann, H., Kapadnis, P. B., Deiters, A., and Chin, J. W. (2009) Genetic encoding and labeling of aliphatic azides and alkynes in recombinant proteins via a pyrrolysyl-tRNA Synthetase/tRNA(CUA) pair and click chemistry. J. Am. Chem. Soc. 131 (25), 8720−8721. (35) Ahmad, S., Gromiha, M., Fawareh, H., and Sarai, A. (2004) ASAView: database and tool for solvent accessibility representation in proteins. BMC Bioinf. 5, 51. (36) Mbua, N. E., Guo, J., Wolfert, M. A., Steet, R., and Boons, G. J. (2011) Strain-promoted alkyne-azide cycloadditions (SPAAC) reveal new features of glycoconjugate biosynthesis. ChemBioChem 12 (12), 1912−1921. (37) Beckley, N. S., Lazzareschi, K. P., Chih, H. W., Sharma, V. K., and Flores, H. L. (2013) Investigation into temperature-induced aggregation of an antibody drug conjugate. Bioconjugate Chem. 24 (10), 1674−1683. (38) Zhang, C., Yao, T., Zheng, Y., Li, Z., Zhang, Q., Zhang, L., and Zhou, D. (2016) Development of next generation adeno-associated viral vectors capable of selective tropism and efficient gene delivery. Biomaterials 80, 134−145.

(6) Adumeau, P., Sharma, S. K., Brent, C., and Zeglis, B. M. (2016) Site-Specifically Labeled Immunoconjugates for Molecular Imaging– Part 1: Cysteine Residues and Glycans. Mol. Imaging Biol. 18 (1), 1− 17. (7) Axup, J. Y., Bajjuri, K. M., Ritland, M., Hutchins, B. M., Kim, C. H., Kazane, S. A., Halder, R., Forsyth, J. S., Santidrian, A. F., Stafin, K., et al. (2012) Synthesis of site-specific antibody-drug conjugates using unnatural amino acids. Proc. Natl. Acad. Sci. U. S. A. 109 (40), 16101− 16106. (8) Sochaj, A. M., Swiderska, K. W., and Otlewski, J. (2015) Current methods for the synthesis of homogeneous antibody-drug conjugates. Biotechnol. Adv. 33 (6), 775−784. (9) Adem, Y. T., Schwarz, K. A., Duenas, E., Patapoff, T. W., Galush, W. J., and Esue, O. (2014) Auristatin antibody drug conjugate physical instability and the role of drug payload. Bioconjugate Chem. 25 (4), 656−664. (10) Voynov, V., Chennamsetty, N., Kayser, V., Wallny, H. J., Helk, B., and Trout, B. L. (2010) Design and application of antibody cysteine variants. Bioconjugate Chem. 21 (2), 385−392. (11) Junutula, J. R., Bhakta, S., Raab, H., Ervin, K. E., Eigenbrot, C., Vandlen, R., Scheller, R. H., and Lowman, H. B. (2008) Rapid identification of reactive cysteine residues for site-specific labeling of antibody-Fabs. J. Immunol. Methods 332 (1−2), 41−52. (12) Lyons, A., King, D. J., Owens, R. J., Yarranton, G. T., Millican, A., Whittle, N. R., and Adair, J. R. (1990) Site-specific attachment to recombinant antibodies via introduced surface cysteine residues. Protein Eng., Des. Sel. 3 (8), 703−708. (13) Shinmi, D., Taguchi, E., Iwano, J., Yamaguchi, T., Masuda, K., Enokizono, J., and Shiraishi, Y. (2016) One-Step Conjugation Method for Site-Specific Antibody-Drug Conjugates through Reactive Cysteine-Engineered Antibodies. Bioconjugate Chem. 27 (5), 1324−1331. (14) Agarwal, P., and Bertozzi, C. R. (2015) Site-specific antibodydrug conjugates: the nexus of bioorthogonal chemistry, protein engineering, and drug development. Bioconjugate Chem. 26 (2), 176−192. (15) Zhou, Q., Stefano, J. E., Manning, C., Kyazike, J., Chen, B., Gianolio, D. A., Park, A., Busch, M., Bird, J., Zheng, X., et al. (2014) Site-specific antibody-drug conjugation through glycoengineering. Bioconjugate Chem. 25 (3), 510−520. (16) Li, X., Fang, T., and Boons, G. J. (2014) Preparation of welldefined antibody-drug conjugates through glycan remodeling and strain-promoted azide-alkyne cycloadditions. Angew. Chem., Int. Ed. 53 (28), 7179−7182. (17) Drake, P. M., Albers, A. E., Baker, J., Banas, S., Barfield, R. M., Bhat, A. S., de Hart, G. W., Garofalo, A. W., Holder, P., Jones, L. C., et al. (2014) Aldehyde tag coupled with HIPS chemistry enables the production of ADCs conjugated site-specifically to different antibody regions with distinct in vivo efficacy and PK outcomes. Bioconjugate Chem. 25 (7), 1331−1341. (18) Dennler, P., Chiotellis, A., Fischer, E., Bregeon, D., Belmant, C., Gauthier, L., Lhospice, F., Romagne, F., and Schibli, R. (2014) Transglutaminase-based chemo-enzymatic conjugation approach yields homogeneous antibody-drug conjugates. Bioconjugate Chem. 25 (3), 569−578. (19) Zeglis, B. M., Davis, C. B., Abdel-Atti, D., Carlin, S. D., Chen, A., Aggeler, R., Agnew, B. J., and Lewis, J. S. (2014) Chemoenzymatic strategy for the synthesis of site-specifically labeled immunoconjugates for multimodal PET and optical imaging. Bioconjugate Chem. 25 (12), 2123−2128. (20) Shen, B. Q., Xu, K., Liu, L., Raab, H., Bhakta, S., Kenrick, M., Parsons-Reponte, K. L., Tien, J., Yu, S. F., Mai, E., et al. (2012) Conjugation site modulates the in vivo stability and therapeutic activity of antibody-drug conjugates. Nat. Biotechnol. 30 (2), 184−189. (21) Jackson, D., Atkinson, J., Guevara, C. I., Zhang, C., Kery, V., Moon, S. J., Virata, C., Yang, P., Lowe, C., Pinkstaff, J., et al. (2014) In vitro and in vivo evaluation of cysteine and site specific conjugated herceptin antibody-drug conjugates. PLoS One 9 (1), e83865. (22) Adumeau, P., Sharma, S. K., Brent, C., and Zeglis, B. M. (2016) Site-Specifically Labeled Immunoconjugates for Molecular Imaging2467

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468

Article

Bioconjugate Chemistry (39) Cook, B. E., Adumeau, P., Membreno, R., Carnazza, K. E., Brand, C., Reiner, T., Agnew, B. J., Lewis, J. S., and Zeglis, B. M. (2016) Pretargeted PET Imaging Using a Site-Specifically Labeled Immunoconjugate. Bioconjugate Chem. 27 (8), 1789−1795. (40) Cooper, M. S., Ma, M. T., Sunassee, K., Shaw, K. P., Williams, J. D., Paul, R. L., Donnelly, P. S., and Blower, P. J. (2012) Comparison of (64)Cu-complexing bifunctional chelators for radioimmunoconjugation: labeling efficiency, specific activity, and in vitro/in vivo stability. Bioconjugate Chem. 23 (5), 1029−1039. (41) Cooper, M. S., Sabbah, E., and Mather, S. J. (2006) Conjugation of chelating agents to proteins and radiolabeling with trivalent metallic isotopes. Nat. Protoc. 1 (1), 314−317.

2468

DOI: 10.1021/acs.bioconjchem.6b00412 Bioconjugate Chem. 2016, 27, 2460−2468